Automated 3D tomographic velocity analysis of residual moveout in prestack depth migrated common image point gathers

Abstract

Residual moveout analysis of prestack depth migrated common image point gathers is used for velocity model building in areas of complex geology. Where the velocity varies smoothly, the model is commonly built from stacking velocity analysis of the residual curvature on the gathers after conversion back to time. Where the velocity varies more rapidly, the model must be built instead by projecting the residual depth errors back over individually traced raypaths in a full tomographic inversion in depth. In this paper we show results of the latter method applied to a Gulf Coast data set. The 3D field example was run in a highly automated fashion. The depth errors (residual moveout) were first picked in batch mode throughout a sparsely sampled volume of common image point gathers and then minimized through a global inversion of all picks at all depths simultaneously. The inverse problem was constrained with preconditioning to solve for the smoothest part of the velocity field first. This automated approach improves turnaround by minimizing human intervention.